summary

Calculates summary statistics.

Description

Calculates summary statistics to provide a quick overview of data.

There are two functions here: the first calculates min, max, mean, median, standard deviation, and does not require any additional toolboxes. The second also calculates quartiles, and the frequency of each data value. This version requires the statistics toolbox.

• Required Products: Statistics Toolbox
• MATLAB release: MATLAB 7.1.0 (R14SP3)

[Duane Hanselman]

What is "the R/S-Plus function of the same name"? Same name as what? What is q1? What is q3? What are the input arguments? You do not describe how to use this function. Since you use varargin, I assume that the user can call the function with an arbitrary number of input arguments. What purpose do those added inputs serve? Where is the unspecified function prctile? If this function requires a specific MATLAB toolbox, please say so. What is or are the outputs of this function? You mention them briefly here on FEX, but a complete description belongs in the help text of the function itself. In its current state this function does not have sufficient value to other users.

[urs (us) schwarz]

potentially useful snipplet for users who need to get a quick and dirty overview of their data in gaussian terms
while it seems to be addressed mainly at users familiar with the r/s-plus environment, i fully agree with prof hanselman's review (in view of ML users)
- needs better help
- MUST state the fact that it requires the STATISTICS TBX!
- -and- a pointer to this tbx's GRPSTATS would be appropriate
us

[John D'Errico]

Long ago I wrote a function like this. I've found it very useful over the years to give a quick complete set of statistics on some variable. (I'd suggest that std should definitely be added to the list of statistics that it computes.) What would it take to make this code useful? Good help. Its not hard to write. Even the best tool is useless without it though.

[Richie Cotton]

Thanks for the criticism, a less rubbish version is on its way. This time it'll have comments and everything.

[Richie Cotton]

Now updated with criticisms accounted for. I know the premise isn't groundbreaking, but I think this is still a useful (ans simple) little tool for getting a feel for a dataset. The code is now fully commented, and there is a version that doesn't require the stats toolbox.

[John D'Errico]

It seems quite reasonable now. As a tool, its a very useful one. I learned long ago that the best way to work (for me) was to verify every step of my analysis as I do the work. Then when my analysis was done, I'd also be confident it was correct. I'd verify that the min and max of my data is realistic, that its order of magnitude is what I would expect, that its general distribution is what I'd expect, and that the size of my array is also correct. (That is the only other thing I'd add. Sorry - I forgot to mention that useful item in my last review.) This is also a good tool when you are debugging code. Finally, its the very first tool I use when I receive some new data.

[ashkan]